Nieman Lab: how tiny newsrooms use AI and what works
What the article is about
Published by Nieman Journalism Lab in July 2026, this piece by Mark Glaser surveys six small news operations — typically one or two staff members covering local beats — about which AI tools they use, what tasks they apply them to, and what practical guidance they would offer other small teams. The organizations include Alaska’s Mat-Su Sentinel, Stumptown Savings in Portland, and Connect Puerto Rico, among others.
Context: small newsrooms, real constraints
The piece operates in a context that is different from typical AI-in-journalism coverage, which tends to focus on large newsrooms with dedicated AI teams and resources. The operations Glaser surveys have no such infrastructure. Their AI adoption decisions are made by the same person filing stories and managing subscriptions. The question is not which AI strategy to pursue but which tools actually reduce workload without creating new problems.
Key takeaways
The range of applications documented is broader than might be expected. Copy editing via specialized tools like Copy Edit Pro (for AP style) handles a task that traditionally required a separate person or process. Headline and wordsmithing assistance functions as a collaborative thesaurus for phrasing alternatives. Research and government data analysis via Perplexity saves hours on sourcing work. Transcription through Otter.ai removes a time-consuming bottleneck in interview-heavy reporting. Visualization and infographics are generated through DALL-E and similar tools for stories that would otherwise go without illustration.
The most technically ambitious use came from Bryan Vance at Stumptown Savings, who built interactive resources — including a grocery price tracker and a farmers market directory — using Claude, and then extended that work to build the publication’s entire WordPress theme. This represents a category of capability that would have required contracting out to a developer a few years ago.
A consistent thread through the accounts is transparency with readers. Several of the journalists describe maintaining explicit public policies disclosing how and when AI is used, treating it as a practice that sustains trust rather than an admission to be managed.
Who it is useful for
Journalists and editorial teams at small to mid-size publications who want concrete examples of AI integration rather than strategic frameworks. The article is also useful for anyone managing a writing or content operation with limited staff who wants to understand which categories of AI assistance have proven genuinely useful at a practical level. The examples are specific enough to adapt directly.